Open AccessDISTRIBUTED CONSENSUS-BASED CALIBRATION OF NETWORKED CONTROL SYSTEMS
Author(s) -
Maja Stanković,
Dragan Antić
Publication year - 2020
Publication title -
facta universitatis. series: automatic control and robotics
Language(s) - English
Resource type - Journals
eISSN - 1820-6425
pISSN - 1820-6417
DOI - 10.22190/fuacr1902095s
Subject(s) - recursion (computer science) , offset (computer science) , convergence (economics) , computer science , rate of convergence , algorithm , control theory (sociology) , control (management) , artificial intelligence , telecommunications , channel (broadcasting) , economics , programming language , economic growth
In this paper a new algorithm for distributed blind macro-calibration of Networked Control Systems is presented. It is assumed that the measured signal is stochastic and unknown. The algorithm is in the form of recursions of gradient type for estimation of the correction parameters for sensor gains and offsets. The recursion for gain correction is autonomous, derived from the measurement increments. The recursion for offset correction is based on differences between local measurements and utilizes the results of gain correction. It is proved that the algorithm provides asymptotic convergence to consensus in the sense that the corrected gains and offsets are equal for all sensors. It is demonstrated that the adopted structure of the algorithm enables obtaining high convergence rate, superior to the algorithms existing in the literature. Simulation results are provided illustrating the proposed algorithm properties.